Charles Yarborough, MD, MPH, FACOEM, FACPM

Transcription

1 Charles Yarborough, MD, MPH, FACOEM, FACPM 1

2 Disclaimers I have not had significant direct funding with any pharmaceutical company for over 20 years. This presentation represents my personal views and not those of a current or past employer. The data is this presentation is either pro forma or in the published literature; there is no proprietary data. 2

3 CER Questions Do we need Comparative Effectiveness Research (CER) to improve plan members health status? Will CER methods be selectively applied to companies data sets? CER methods be important for assessing population impacts of alternative healthcare approaches? (an example) 3

4 Per Member Net Payments Population Health Imperative Pro Forma data $400 Healthy/ Non-Users 35% Prevalence 5% Costs $1,500 Stable 30% Prevalence 10% Costs $4,000 At Risk 20% Prevalence 25% Costs Illness Burden $11,000 Struggling 10% Prevalence 30% Costs $42,000 In Crisis 5% Prevalence 30% Costs If you do nothing, 2% of your population gets sicker every year. * * Dr. Dee Edington, Director, Health Management; Research Center, University of Michigan 4

5 Population Health as Strategic Component Investments in health care benefits Line of Sight Business Value Actionable metrics that measure and monitor: Financial Management Population Health Management Stakeholder Engagement Supplier Performance Management 5

6 Potential for Secondary Analyses of Existing Clinical Datasets for CER Typical CER Estimating incidence and prevalence Estimating treatment needs Developing health policy Testing clinical hypotheses Performing meta-analyses Innovative challenges* (examples) Cross-design synthesis to standardize and compare clinical data collected by different methods Evaluation of new statistical models and methods on treatment effectiveness outcomes *Ref.: NIH Challenge Grant 05-AA-101 (high priority) 7

7 NIH Challenge Grants for CER Prevention and Risk Factor Reduction Strategies for Disabilities (05-AG-102*) Comparative Effectiveness Studies of Non- Pharmacological Treatments for Chronic Low Back Pain (05-AT-101*) CER on Cancer Screening (05-CA-102*) Cost-Effectiveness of Patient Navigation (05-CA-103*) Understanding the Effects of Bariatric Surgery on Type 2 Diabetes and Cardiovascular Risk Factors (05-DK-101*) Support Pilot CER Projects in Community Settings (05-RR-102*) 8

8 Health of Individuals and Populations 9

9 Population Impact Assessments Ask the question -- make the options explicit Collect data Relevant population denominators Prevalence and current practice Estimated data on baseline risk of identified outcomes Library of evidence for risks (Relative Risks and Relative Risk Ratios ) Calculate impact -- population impact measures Understand -- values, training, culture, receptivity Use -- implement results in prioritizing services using change and knowledge management 10

10 Population Impact Numbers Need: Public health counterpart to evidence based medicine Solution: Population impact measures to use evidence that is combined with collected data to provide local context to measure of risk and benefit, and to support health policy decision making Types: Eliminating a risk factor (PIN-ER-t) and the number of events prevented by the intervention in your population (NEPP) Ref.: Heller RF, et al Brit Med J

11 Total Population Population Impact * Includes embedded numbers # Diseased* # Treated* # Events Prevented by Intervention* Adapted from R. Heller Evidence for Population Health (2005), Oxford U. Press 12

12 Atorvastatin vs. Simvastatin Study* Case-referent design Full time employees from 23 companies with 2 years continuous postindex enrollment Matching (13,584 in each group) Initial drug dose Baseline inpatient CVD events Average wage Propensity score Primary study outcomes collected for 2 years Rate of inpatient CVD events Total costs to employers *Simpson RJ, et al. 2009; Mayo Clin Proc 84(12): (authors funded by Pfizer) 13

13 Δ # cardiovascular events in 2 years Impact at 40% Prevalence by Rx Mix and Compliance Atorvastatin Simvastatin Ratio 14

14 Δ # cardiovascular events in 2 years Impact at 20% Prevalence by Rx Mix and Compliance Atorvastatin Simvastatin Ratio 15

15 Δ $ (2006) in millions Added Cost for Index Drug Depending on Prevalence and Rx Mix Atorvastatin Simvastatin Ratio Atorvastatin $946 mean; Simvastatin $489 mean. 16

16 Summary We need CER studies to enhance decision making for improving plan members health status. CER methods can be selectively applied to and supplemented by companies data sets. CER methods be important for assessing population impacts of alternative healthcare approaches. 17

17 Thank You 18